| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | pgsql-hackers(at)lists(dot)postgresql(dot)org |
| Subject: | ExecAggTransReparent is underdocumented and badly named |
| Date: | 2023-04-19 18:50:51 |
| Message-ID: | 3004282.1681930251@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
I wondered at [1] why ExecAggTransReparent doesn't do what you'd
guess from the name, that is reparent a supplied R/W expanded object
under the aggcontext even if it's not there already. I tried to
make it do that, and after I'd finished bandaging my wounds,
I wrote some commentary explaining why that won't work. (Possibly
we knew this at some point, but if so we sure failed to provide
documentation about it.)
I think we should at least apply the attached commentary-only
patch. I wonder though if we should change the name of this
function, and if so to what. Maybe ExecAggCopyTransValue?
regards, tom lane
[1] https://www.postgresql.org/message-id/2199319.1681662388%40sss.pgh.pa.us
| Attachment | Content-Type | Size |
|---|---|---|
| document-ExecAggTransReparent-better.patch | text/x-diff | 3.9 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andres Freund | 2023-04-19 19:20:51 | Re: Wrong results from Parallel Hash Full Join |
| Previous Message | Peter Geoghegan | 2023-04-19 18:38:48 | Re: Should we put command options in alphabetical order in the doc? |